Seminar Series
The AI Sentience Scholars (AISS) Program , an initiative from Neuromatch, supports early-career researchers exploring questions at the intersection of AI, consciousness, and ethics.
AISS is a 6-month, part-time, remote research and training program in which scholars develop mentored research projects while engaging with conceptual foundations, ethical frameworks, and the broader implications of advanced AI systems. The program approaches these topics from a neutral, inquisitive, and critically grounded perspective, prioritizing empirical rigor and open inquiry.
About the Seminar Series
The AISS Seminar Series exposes scholars to diverse perspectives across academia, industry, policy, and applied research related to AI sentience and the broader study of intelligent systems. The series aims to foster interdisciplinary dialogue and critical reflection by bringing together researchers, practitioners, and thought leaders working at the intersection of AI, cognitive science, neuroscience, philosophy, governance, and society.
Sessions are open to the broader community and feature invited external speakers, alongside mentors and collaborators connected to the program.
We warmly invite mentors and community members to volunteer as speakers, suggest external speakers, or act as session hosts. If you would like to give or propose a seminar, please contact the program team . Topics may include scientific advances, research methods, interdisciplinary perspectives, career paths, or lessons learned from practice.
Format: ~30 min talk + ~20 min discussion and Q&AAudience: Interdisciplinary scholars, mentors, and community members
Regular slot: Wednesdays, usually 15:00 UTC, approximately every 3 weeks (flexible depending on speaker availability)
Jul 22, 2026
15:00 UTC
Kim Stachenfeld (Google DeepMind, Columbia University)
Discovering Interpretable Symbolic Models of Human and Animal Behavior with LLMs
Aug 19, 2026
17:00 UTC
Winnie Street & Geoff Keeling (Google Research, University of London)
Emerging Questions in AI Welfare
Sep 9, 2026
15:00 UTC
Susan Schneider & Mark Bailey (Florida Atlantic University)
Sep 30, 2026
15:00 UTC
Robert Long (Eleos AI)
Oct 21, 2026
15:00 UTC
Michael J Tarr (Carnegie Mellon University)
Nov 11, 2026
Jean-Remi King (CNRS, Meta) (tbc)
Nov 25, 2026
Adina Roskies (University of California, Santa Barbara) (tbc)
.png?apiToken=eyJhbGciOiJIUzI1NiIsImtpZCI6IjIwMjMtMDUtMDQifQ.eyJzY29wZSI6Im5vdGUtZXhwb3J0IiwibmlkIjoiS1JGZ3FEQkxvLVlNcmciLCJpYXQiOjE3ODc3MDA3OTgsImlzcyI6Imh0dHBzOi8vc2xpdGUuY29tIiwianRpIjoiY3o4anJ4TzNGcnJkUEwiLCJleHAiOjE3OTAyOTI3OTh9.Xk4z2VLWqVcc_NdxqyCrtpXYNkCsDndA96itEh-CXco)
Discovering Interpretable Symbolic Models of Human and Animal Behavior with LLMs
Presenter: Kim Stachenfeld from Google DeepMind, Columbia University
Abstract: Symbolic models play a key role in neuroscience and psychology, expressing computationally precise hypotheses about how the brain implements a cognitive process. Identifying an appropriate model typically requires a great deal of effort and ingenuity on the part of a human scientist. This talk covers DataDIVER, a recently developed technique for automatically discovering interpretable symbolic models that accurately capture human and animal learning. DataDIVER leverages the ability of large language models to automatically generate code to explore a vast space of candidate models, and returns a set of models that each strike different balances between quality-of-fit and simplicity. The approach is applied to a number of datasets containing learning behavior from a range of species and reward-guided learning tasks. The best-fitting programs match the quality-of-fit of "blackbox" neural network models. The remaining spectrum of programs surfaces meaningfully novel insights in a more accessible way, with the simplest models shedding light on the basic organization of learning behavior, and more complex programs revealing more detailed structure. Some of these discovered learning mechanisms suggested the presence of previously unknown patterns, verified by reexamining the behavioral data. Broadly, these results show that AI tools can be used not just to predict data but also to explain it.
Presenter information:
Kim Stachenfeld is a Research Scientist at Google DeepMind in NYC and Affiliate Faculty at the Center for Theoretical Neuroscience at Columbia University . Her research covers topics in Neuroscience and AI. On the Neuroscience side, she studies how animals build and use models of their world that support memory and prediction. On the Machine Learning side, she works on implementing these cognitive functions in deep learning models. Kim’s work has been featured in The Atlantic , Quanta Magazine , Nautilu s, and MIT Technology Review . In 2019, she was named one of MIT Tech Review’s Innovators under 35 for her work on predictive representations in hippocampus.
Resources:
Emerging Questions in AI Welfare
Presenters: Winnie Street and Geoff Keeling from Google Research, University of London
Abstract: In this talk we investigate whether artificial intelligence (AI) systems could ever be welfare subjects, understood as entities for which things can go better or worse. Some people argue that AIs could plausibly have or soon have features like consciousness, agency, and the capacity for social relationships, which could in principle provide a basis for AI welfare. These arguments have massive significance for the societal conversation on AI, raising profound ethical and political questions about what if anything we owe to these new technologies. We here provide the philosophical groundwork for a scientific, philosophical, and ultimately democratic inquiry into the potential for AI welfare, addressing key questions that cut across different arguments: what welfare is, how to interpret behavioural evidence of AI welfare, what kinds of entities might qualify as candidate AI welfare subjects, the potential grounds for welfare in AI, and the practical ethical challenges that arise from our uncertainty.
Presenter information:
Winnie Street is a Senior Research Scientist at Google and a Fellow at the Institute of Philosophy in the School of Advanced Study, University of London. Her research combines philosophical and empirical approaches to questions concerning the cognitive capacities and ethical significance of frontier AI systems, including whether AI systems could ever be conscious, whether they could be moral patients, and how we should understand the relationships that people build with them.
Geoff Keeling , PhD is a Staff Research Scientist at Google, a Fellow at the Institute of Philosophy in the School of Advanced Study, University of London, and an Associate Fellow at the Leverhulme Centre for the Future of Intelligence, University of Cambridge. His work focuses on the ethics and cognitive science of frontier artificial intelligence systems including disputes about alignment, manipulation, trust, digital minds and human-AI relationships.
Date: 15 September 2026
Time: 15:00 UTC
Register for free: https://us06web.zoom.us/meeting/register/pOUPl7GQTbK8zQ5fyMR_7Q#/registration 30-min talk + discussion & Q&A
Presenter information: Susan Schneider & Mark Bailey (Florida Atlantic University (FAU), Center for the Future of AI, Mind and Society)